Pakistan Journal of Meteorology Vol. 10, Issue 19: July, 2013 Mean Temperature Projection by PRECIS over Pakistan for 2001-2010 Ali, G. 1, 2 , W. Iqbal 2 , G. Rasul 2 Abstract The regional climate model PRECIS (Providing Regional Climate for Impact Studies) of the Headley Centre, Met Office UK, was used to simulate temperature along with other parameters at 0.22 ° × 0.22 ° (about 25 km × 25 km) resolution for Pakistan. The PRECIS output mean temperature was compared with two datasets; (1) a published dataset of CRU (Climate Research Unit) developed by the University of East Anglia, which is a dataset prepared from the real time data of station observatories all over the world, and (2) real time temperature data of the observatories of Pakistan Meteorological Department (PMD). While there is a large scale agreement between these datasets, considerable regional differences also exist. During the summer season (JJA, i.e. June, July and August) the PRECIS is warm bias in Punjab, southern Khyber Pakhtoonkhwa, northern Sindh and the adjoining areas, while this trend shrinks to a smaller region during the winter season of DJF (i.e. December, January and February) as compare to the observed temperature data sets (i.e. both CRU and PMD observed temperature data). Overall the PRECIS predicted temperature is in good agreement with CRU and PMD observed temperature in the lower plane areas of the country with slight differences, whereas in the northern hilly areas of complex topography there exist large biases between these datasets (i.e. PRECIS and the two observed datasets). Thus the model behavior is rather a chaotic one in the regions with latitude greater than 30 ° (more precisely in the regions with latitudes greater than 33 ° ). The correlation between PRECIS and PMD observed temperature in the regions with latitude smaller than 30 ° is better than the correlation between them in the region with latitude greater than 30 ° . Key Words: Regional Climate Model, Time Series, Correlations, topography. Introduction General Circulation Models (GCMs) have assisted very well in understanding of anthropogenic global climate change and this has helped to develop better mitigation strategies. But the GCM results are usually of course resolution and the climate change projection based on these GCMs lacking local details that are important for impact studies on national and regional levels. Thus at the regional scale the relevant and targeted projection for regional climate change is becoming more and more imperative. Demand for regional climate change scenarios has generated increased interest in downscaling of global climate model simulations. Different downscaling methods are in use nowadays among which the most important are statistical downscaling and dynamical downscaling. The latter one is carried out using regional climate models (RCM) (Lincoln M. Alves & José Marengo, 2010). The RCM PRECIS (Providing Regional Climate for Impact Studies) is an atmospheric and land surface model of limited area and high resolution. It is developed by Hadley Center at the UK Met office for climate simulation and research. This model generates high resolution climate change information and it can be applied to any area of the globe for the development of climate change scenarios for any region of the world. The large scale predictions made by a GCM are supplied to PRECIS which add detail information to these predictions. It has the provision to include sulfur cycle with the ability to generate output of about 150 parameters. PRECIS is based on the atmospheric component of HadCM3 climate model (Gordon et al., 2000) and it is described extensively in Jones et al. (2004). The atmospheric dynamics module of PRECIS is a hydrostatic version of the full primitive equations and uses a regular latitude-longitude grid in the horizontal and a hybrid vertical coordinate. There are 19 levels from the ground up to 0.5 hPa and it can 1 gohar.met@gmail.com 2 Pakistan Meteorological Department, Pitras Bukhari Road, Sector H-8/2, Islamabad. 31